resched_init_time = af.humanToSeconds(t)
URIDs = af.get_URID_Bus(fullSchedule_windows, broken_run,
resched_init_time)
else:
for i in range(len(individual_requests)):
if individual_requests[i] not in list(
set(fullSchedule_windows.BookingId.tolist())):
print(
'ERROR 405: You have entered BookingIds not present in the schedule!'
)
flag = 405
return flag
URIDs = af.get_URID_BookingIds(fullSchedule_windows,
individual_requests)
if not URIDs:
print(
'ERROR 406: There are no people to reschedule on bus {0} at time {1}'
.format(broken_run, resched_init_time))
flag = 406
return flag
# for each URID we find the bus runs to check through a radius elimination.
# for each URID for each run we then want to check the capacity in the given time
# window and return the URID with updated insert points. This URID with updated
# insert points is fed to the feasibilty function, which we ultimately want to return
# a minimum cost run for the URID and that run updated with the new URID slotted in.
taxi_costs = []
delay_costs = []
def busReschedule_run(schedule_filename,
accesskey,
secretkey,
broken_run,
path_to_outdir = af.os.path.join(af.os.getcwd(),'data'),
resched_init_time = None,
bookingid = None,
windows = 1800.,
radius = 3.):
'''
schedule_filename (str): name of file to be used if you want to test a DEMO file. Must be a single day, QC'ed file.
accesskey (str): AWS_ACCESS_KEY
secretkey (str): AWS_SECRET_KEY
bookingid (?): list of BookingIds to be rescheduled (comma separated?)
broken_run (str): name of broken run.
windows (float): size of time windows, in seconds
resched_init_time (int): seconds in day at which rescheduling can occur.
path_to_outdir (str): path to directory where you would like to store all output files.
radius (float): number of miles to search for close buses. Will be shrunk by increments of 1 if
number of nearby buses is > 30.
'''
flag = 200 #400's are bad, 200 is good.
if not af.os.path.exists(path_to_outdir):
path_to_outdir = af.os.path.join(af.os.getcwd(),'data')
#get rescheduling data from the webapp/data directory
if schedule_filename is not None:
if af.os.path.isfile(af.os.path.join(path_to_outdir, schedule_filename)):
fullSchedule = af.pd.DataFrame.from_csv(schedule_filename, header=0, sep=',', index_col = False)
else:
print('ERROR 401: Demo file not found!')
flag = 401
return flag
if accesskey and secretkey is not None:
#AWS_ACCESS_KEY = raw_input("Please enter AWS access key: ")
#AWS_SECRET_KEY = raw_input("Please enter AWS secret key: ")
try:
fullSchedule = af.s3_data_acquire(accesskey, secretkey, path_to_outdir, qc_file_name = 'qc_streaming.csv')
if type(fullSchedule) == int:
print("ERROR 402: formatting error in streaming data.")
flag = 402
return flag
except IOError: #is this the right error if s3_data_acquire fails?
print('ERROR 403: Could not access streaming data!')
flag = 403
return flag
#Determine broken run number, or get list of unhandled requests.
if broken_run is not None:
case = 'BROKEN_RUN'
elif bookingid is not None:
case = 'INDIVIDUAL_REQUESTS'
individual_requests = list(bookingid)
# this simply returns full schedule with time windows at the moment
sched_obj = af.aTWC.TimeWindowsCapacity(fullSchedule)
fullSchedule_windows = sched_obj.addtoRun_TimeCapacity(windows)
fS_w_copy = fullSchedule_windows.copy()
#this gets us all the URIDs for the broken run given the initial rescheduling time
#OR it will get us URIDs given specific bookingIds to be rescheduled
if case == 'BROKEN_RUN':
if broken_run not in list(set(fullSchedule_windows.Run.tolist())):
print('ERROR 404: Run number is not scheduled for today!')
flag = 404
return flag
if resched_init_time is not None:
resched_init_time = af.humanToSeconds(resched_init_time)
if resched_init_time is None:
t = af.datetime.datetime.now()
hr = str(t.hour)
if t.minute < 10:
m = str(0)+str(t.minute)
else:
m = str(t.minute)
t = hr+':'+m
resched_init_time = af.humanToSeconds(t)
URIDs = af.get_URID_Bus(fullSchedule_windows, broken_run, resched_init_time)
else:
for i in range(len(individual_requests)):
if individual_requests[i] not in list(set(fullSchedule_windows.BookingId.tolist())):
print('ERROR 405: You have entered BookingIds not present in the schedule!')
flag = 405
return flag
URIDs = af.get_URID_BookingIds(fullSchedule_windows, individual_requests)
if not URIDs:
print('ERROR 406: There are no people to reschedule on bus {0} at time {1}'.format(broken_run, resched_init_time))
flag = 406
return flag
# for each URID we find the bus runs to check through a radius elimination.
# for each URID for each run we then want to check the capacity in the given time
# window and return the URID with updated insert points. This URID with updated
# insert points is fed to the feasibilty function, which we ultimately want to return
# a minimum cost run for the URID and that run updated with the new URID slotted in.
taxi_costs = []
delay_costs = []
best_buses = []
for i in range(len(URIDs)):
print('Rescheduling URID {0}'.format(i))
busRuns_tocheck = af.radius_Elimination(fullSchedule_windows, URIDs[i], radius=radius)
insert_stats = []
#iterate over all runs, find best one!
for run in busRuns_tocheck:
this_run = fullSchedule_windows[fullSchedule_windows['Run']==run]
capacity_obj = af.checkCap.CapacityInsertPts(this_run)
#Kristen's capacity checker:
URIDs[i].PickupInsert, URIDs[i].DropoffInsert = capacity_obj.return_inserts(URIDs[i])
# IF THERE'S ROOM: TEST FEASIBILITY
if not af.np.isnan(URIDs[i].PickupInsert):
URIDs[i].PickupStart = URIDs[i].PickupInsert
URIDs[i].DropoffStart = URIDs[i].DropoffInsert
runSchedule = af.get_busRuns(fullSchedule_windows, run, None)
print('Testing feasibility for run ' + run)
brokenwindows_dict =af.insertFeasibility(runSchedule, URIDs[i])
if not brokenwindows_dict:
print('Run {0} infeasible without moving the return-to-garage row.'.format(run))
else:
insert_stats.append(brokenwindows_dict)
#ASSEMBLE and ORDER transit options.
if insert_stats:
#ORDER buses by lowest additional lag time, i.e. total_lag, and sequentially add total_lag's
ordered_inserts = sorted(insert_stats, key = af.operator.itemgetter('additional_time'))
popme = []
for k in range(len(ordered_inserts)):
if ordered_inserts[k]['RunID']==broken_run:
popme.append(k)
if popme:
ordered_inserts.pop(popme)
delay_costs.append(ordered_inserts[0]['additional_time'][0]*(48.09/3600)) #total dollars
best_buses.append(ordered_inserts[0]['RunID'])
#CALCULATE taxi cost
taxi_costs.append(af.taxi(URIDs[i]))
#WRITE information about best insertions to text file
if len(ordered_inserts) >= 3:
af.write_insert_data(URIDs[i], ordered_inserts[0:3],
path_to_outdir, taxi_costs[i])
else:
af.write_insert_data(URIDs[i], ordered_inserts[0:],
path_to_outdir, taxi_costs[i])
#UPDATE whole day's schedule:
fullSchedule_windows = af.day_schedule_Update(data = fullSchedule_windows, top_Feasibility = ordered_inserts[0], URID = URIDs[i])
sched_obj_update = af.aTWC.TimeWindowsCapacity(fullSchedule_windows)
fullSchedule_windows = sched_obj_update.add_Capacity(update = True)
#SAVE just the updated run for each URID
fullSchedule_windows[fullSchedule_windows['Run'] == ordered_inserts[0]['RunID']].to_csv(af.os.path.join(path_to_outdir, str(str(int(URIDs[i].BookingId))+'_schedule.csv')), index = False)
else:
delay_costs.append(400000)
taxi_costs.append(af.taxi(URIDs[i]))
af.write_insert_data(URIDs[i], None, path_to_outdir, taxi_costs[i])
best_buses.append('NA')
#WRITE csv of PREFERRED OPTIONS:
if case == 'BROKEN_RUN':
nrun_cost = af.newBusRun_cost(af.get_busRuns(fS_w_copy, broken_run, URIDs[0]), provider = 6)
else:
nrun_cost = None
pref_opt = af.preferred_options(URIDs, best_buses, delay_costs, taxi_costs, nrun_cost)
pref_opt.to_csv(af.os.path.join(path_to_outdir, 'preferred_options.csv'), index = False)
return flag
